© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 4302
    Prev
    Next
    INDEX
    Explanations

    various forms of "single" and "multiple" in scientific contexts, particularly related to substances, conditions, or methodologies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    scientific terminology related to measurement, experimental methodology, and statistical language.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi

    descriptive terms that modify or classify other words, such as "single," "common," "real-time," "large-scale," and "short term."

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    adjectival modifiers—especially hyphenated or degree/quantity descriptors—used before nouns.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    words that contrast or distinguish items, particularly adjectives describing differences in scale, scope, or type (such as "different", "single", "double", "main", "common", "large-scale", "short term").

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    The
    -0.48
    I
    -0.46
    1
    -0.42
    M
    -0.41
    You
    -0.40
    N
    -0.40
    In
    -0.39
     that
    -0.39
    S
    -0.39
    A
    -0.39
    POSITIVE LOGITS
    <unused42>
    1.21
    <unused41>
    1.21
    <unused16>
    1.21
    <unused28>
    1.21
    <unused3>
    1.21
    [@BOS@]
    1.21
    <unused8>
    1.21
    <unused14>
    1.21
    <unused43>
    1.21
    <unused47>
    1.21
    Activations Density 1.055%

    No Known Activations